Clinical utility of three-dimensional magnetic resonance imaging using a zero-echo-time sequence in endoscopic endonasal transsphenoidal surgery

Background: Recognizing the anatomical orientation surrounding the sellar floor is crucial in endoscopic endonasal transsphenoidal surgery (ETSS). Zero-echo-time (ZTE) sequences were recently suggested for a new bone identification technique on magnetic resonance imaging (MRI). This study aimed to evaluate the clinical usefulness of three-dimensional (3D)-ZTE-based MRI models in providing anatomical guidance for ETSS. Methods: ZTE-based MRI and magnetic resonance angiography (MRA) data from 15 consecutive patients with pituitary tumor treated between September 2018 and May 2019 were used to create 3D-MRI models. From these, the architecture surrounding the sellar floor, particularly anatomical relationships between tumors and internal carotid arteries (ICAs), was visualized to preoperatively plan surgical procedures. In addition, 3D-ZTE-based MRI models were compared to actual surgical views during ETSS to evaluate model applicability. Results: These 3D-ZTE-based MRI models clearly demonstrated the morphology of the sellar floor and matched well with intraoperative views, including pituitary tumor, by successively eliminating sphenoidal structures. The models also permitted determination of the maximum marginal line of the opening of the sellar floor by presenting vital structures such as ICAs and tumors. With such 3D-MRI models, the surgeon could access the intracranial area through the sellar floor more safely, and resect the pituitary tumor maximally without complications. Conclusions: Our 3D-MRI models based on ZTE sequences allowed distinct visualization of vital structures and pituitary tumor around the sellar floor. This new method using 3D-ZTE-based MRI models showed low invasiveness for patients and was useful in preoperative planning for ETSS, facilitating maximum tumor resection without complications.

Endoscopic endonasal transsphenoidal surgery (ETSS) has become increasingly used to access pituitary tumors and other midline skull-based lesions [1,2]. There are several technical advantages to ETSS, such as wider, multidirectional views of the operative field.
Nevertheless, there is a significant risk that surrounding vital structures could be injured, particularly when there is extensive tumor invasion into the sphenoid sinus and involvement of the internal carotid arteries (ICAs) [3]. In such cases, it is important to accurately identify the positional relationships of vital structures preoperatively, especially when the surgeon lacks complete familiarity with the surgical field of ETSS. We have proposed the usefulness of three-dimensional (3D)-computed tomography (CT)-based models or reconstructed 3D-CT/magnetic resonance imaging (MRI) fusion models for obtaining preoperative orientation and providing a road map to ETSS (Fig. 1A, B) [4,5].
However, major obstacles exist to acquiring CT images for creating these 3D models, such as excessive radiation exposure and the necessity for administration of iodocontrast media to depict bilateral ICAs. Recently, zero-echo-time (ZTE)-based segmentation methods on MRI have been suggested as a new bone-identification technique [6][7][8][9]. ZTE was designed to achieve signals from cortical bone for tissue segmentation, and as such can be used to incorporate the bone in MR-based attenuation correction [6][7][8][9]. In addition, magnetic resonance angiography (MRA) techniques have improved markedly and we have been able to easily and clearly evaluate the vasculature without using iodocontrast medium [10]. We therefore created new 3D-MRI models combining ZTE-based MRI to depict the bone of the skull base and MRA to visualize ICAs, which can demonstrate the positional relationship between the pituitary tumor and surrounding sellar structures (Fig.   1C). The present study describes details of this reconstruction process, and reports that our 3D-ZTE-based MRI models show low invasiveness for patients, and high utility in obtaining preoperative orientation regarding vital structures of the sphenoid sinus and providing a road map to ETSS.

Methods
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
The present study was approved by the local ethics committee for clinical research.

Experimental design
In this study, MRI was performed for 15 consecutive cases of endoscopic endonasal transsphenoidal pituitary surgery for pituitary tumor in our hospital between September 2018 and May 2019 ( Table 1). The 3D-MRI-based models were created by the radiologist (Taichi Furumochi) before surgery for each patient. All patients were operated on using these 3D-MRI models in addition to the navigation system (StealthStation ® ; Medtronic, Minneapolis, MN) and an indocyanine green (ICG) endoscope (Karl Storz, Tuttlingen).
Details of these methods are described in the paragraphs that follow. Informed consent was obtained from all individual participants enrolled in the study, including for the surgical procedure and potential risks of ETSS.

Constructed 3D-ZTE based MRI models.
MRI and MRA were performed before surgery using a 3.0-T whole-body MR scanner (General Electric (GE) Healthcare, Waukesha, WI) and an eight-channel phased-array head coil. For each patient, a high-resolution anatomic data set was scanned using 3D spoiled A 3D Advantage Workstation Volume Share 4 (GE Healthcare) was used to process the acquired MRI data, according to our previous reports [4,5].
Step 1 involves rendering volume and defining opacity -250 to +600 as the skin parameter. Step

Statistical methods
Statistical analysis was performed using Fisher's exact test for categorical variables and analysis of variance for continuous variables. Two-tailed tests were performed for each scenario, and the significance level was set at P < 0.05. All analyses were performed using Office Excel 2016 software (Microsoft, Redmond, WA).
We were able to obtain ZTE-based MRI data (proton density image) for all patients ( Fig.   2A). After acquiring the proton density images, we converted these data into CT-like ZTE images to be able to recognize bone (Fig. 2B). In all patients, CT-like bone depiction was identified clearly surrounding the sellar floor in the sphenoid sinus (Fig. 3). In addition, we successfully created 3D-MRI models by combining these ZTE-based MRI data set and acquired MRA images (Figs. 1C, 4, 5). In these models, critical structures such as the ICAs, sellar floor and pituitary tumor were completely reconstructed, and spatial relationships were better visualized by successively deleting images of adjacent bony structure, even if ICAs were involved in the tumor. Furthermore, we clearly recognized the positional relationship between the ICAs and invasive tumor. The 3D-MRI model was able to be viewed in the operating room during surgery as multislice presentations on a computer monitor. Compared to existing 3D-CT-based models or 3D-CT/MRI fusion models, no inferiority was encountered in grasping the anatomical positions surrounding the sellar floor for anatomical guidance in ETSS (Fig. 1), and this reconstruction model had the advantage of needing no radiation exposure or iodine contrast agent.

Case 1
An 80-year-old man complained of headache and deterioration of visual function.
Gadolinium (Gd)-enhanced MRI showed a macroadenoma invading into the right cavernous sinus (Knops classification: grade 3) (Fig. 4A, B). A preoperative 3D-ZTE-based MRI model demonstrated anatomical landmarks within the sphenoid sinus, including bony prominences of the ICAs and pituitary tumor (Fig. 4C). Under the guidance of this 3D-MRI model, we made a preoperative plan for the extent of opening of the sella and carried out ETSS, with modeled images appearing almost the same as the actual intraoperative views (Fig. 4D). In addition, the ICG endoscope allowed visualization of the bilateral ICAs 10 s after ICG flushing in accordance with 3D-MRI model (Fig. 4E). The tumor was totally removed without any complications and the patient showed full recovery of visual acuity and fields.

Case 2
A 52-year-old woman visited our department after demonstrating gradually worsening visual field deficits. Laboratory studies showed a high plasma level of PRL (74.7 pg/ml).

Coronal and sagittal Gd-enhanced MRI showed a macroadenoma involving bilateral ICAs
with invasion into the sphenoid sinus and destruction of the sellar floor (Fig. 5A, B). A preoperative 3D-ZTE-based MRI model clearly revealed the location of bilateral ICAs by successively eliminating bony structures (Fig. 5D), and we identified the exact location of ICAs buried in the tumor (Fig. 5C). As a result, we safely achieved maximum resection of the tumor without injuring the ICAs.

Discussion
ETSS has been found to be highly effective for resecting pituitary tumors with supraand/or infrasellar extension, including lesions of the midline skull base [1,2]. However, the disadvantage of this approach is that, with wide opening of the sellar floor, there is a risk that the surgeon may come extremely close to the ICA. To avoid ICA injury, a 3D-CTbased model or a 3D-CT/MRI fusion model that helps clarify the anatomy of the nasal cavity and paranasal sinuses has been proposed and adopted [4,5]. These models provide clearer orientation to the surgeon, which is critical when performing ETSS. Furthermore, preoperative surgical planning can be performed with optimization of these models according to the individual patient's anatomy [4,5]. However, creation of both of these 3D-CT-based models and 3D-CT/MRI fusion models also has several disadvantages, including excessive radiation exposure for patients from CT and the need to use iodine contrast agents to visualize the ICA.
Generally, MRI is a noninvasive and essential diagnostic technique because of its inherent advantage of obtaining excellent soft-tissue contrast and high resolution of anatomical detail in the body without radiation exposure [9,11]. However, this modality is considered inappropriate for depicting cortical bone structures because of the low proton density and very short T2 relaxation time [9,12]. The current gold standard imaging technique to reveal bone structures is CT. However, a new bone identification technique has recently been published, based on 3D radial ZTE imaging on MRI [8,13,14]. In particular, GE Healthcare has developed an investigational work-in-progress (WIP) MR research package called ZTE for bone imaging, consisting of a pulse sequence technique designed to image cortical bone surfaces [3,4]. The ZTE sequence is a 3D radial sequence used for silent MRI, and the extremely short effective TE allows detection of the shortest T2 tissues, including cortical bone [6,9]. To date, such MR bone imaging using ZTE sequences has been applied to positron emission tomography (PET)/MRI attenuation correction from the perspective of the technical approach in the literature and clinical diagnostic use of osseous shoulder or skull bone imaging [6,9,13]. Above all, in neuroradiology, ZTE skull MRI has been gradually introduced to evaluate skull lesions in patients with head trauma [9]. This ZTE sequence has the advantage of being able to visualize bones without radiation exposure, unlike CT. On the other hand, several problems have been reported for ZTE sequences [6]. In previous reports, the boundary of bone and fluid as found in the inner ear was proposed to be a major problem for ZTE. Furthermore, accuracy was reported to be insufficient at the base of the skull bone or around the paranasal sinuses [6]. In particular, the mastoid air cells, despite also being complex anatomical structures, were found to pose fewer issues. In the present study, we applied the ZTE sequence to obtain images of the skull bones of patients with pituitary tumor. Certainly, according to previous reports, bone depiction using the ZTE sequence is inferior to that using CT

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors Availability of data and materials

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download. Table1.pdf